The taxonomy of Paracoccus denitrificans and related bacteria is discussed. Evidence is given which shows that the physiological differences between P. denitrificans and Thiosphaera pantotropha are less fundamental than previously thought. A proposal to consider a species P. pantotropha is mentioned. The properties of the denitrifying enzymes and the genes involved in their formation in P. denitrificans is discussed. The synthesis of the membrane-bound-nitrate reductase is regulated by FNR, that of the nitrite- and nitric oxide reductase by NNR. Evidence is given that FNR acts as a redox sensor rather than an oxygen sensor. The occurrence of aerobic denitrification and coupled heterotrophic nitrification-denitrification in the original strain of Thiosphaera pantotropha are explained by a limiting respiratory activity which activates FNR. Aerobic denitrification leads to a lower growth yield and an increase in mumax in batch culture when a limiting respiratory activity is assumed and when excess substrate is present. Coupled heterotrophic nitrification-denitrification gives a smaller increase in mumax and a more drastic reduction in yield. Both processes are thus advantageous to the organism. In a chemostat with limiting substrate these processes are disadvantageous. T. pantotropha has lost the ability for aerobic denitrification during extended cultivation. Possibly the substrate concentration was limiting during extended cultivation giving a selective advantage to variants which have lost these properties. The calculations predict that P. denitrificans should be able to grow chemolithotrophically with hydroxylamine.